Re: QTEMP not in library list

["Dave Kahn" <dkahn400@xxxxxxxxxxxxxx>]

On 14/01/2008, Aaron Bartell <albartell@xxxxxxxxx> wrote:
> Scott summarized his viewpoints with the following, which my statement came
> from...
>
> " Let's face it, folks... the only reason QTEMP needs to be in the library
> list (aside from adding it on-the-fly for testing code) is because of
> all the poorly written software out there.  In this market space, poorly
> written software absolutely ABOUNDS -- or, at least, that's been my
> experience."
>
> So maybe I should instead respond with this:  Scott, do you think QTEMP
> should NOT be used for temporary objects that a programmer wants to have
> auto-cleaned-up at the end of the job?  I thought there were good points on
> why a vendor wouldn't want to use QTEMP.  Many times I rely on QTEMP when I
> am doing professional services contracts, and so far it has worked quite
> well for me.  I would chalk QTEMP usage up to "it depends" and "use it to
> your benefit, but understand the drawbacks".
>
> Thoughts?

Scott does not appear to be saying you should never use QTEMP. What I
think he's saying is that software should not depend on the assumption
that QTEMP will be in the library list or, if it is, that it will be
at a particular position within it. Not depending on those assumptions
seems pretty sound to me.